Another false positive
defineindex=: 3 : 0
y=. <;._2 (0 : 0)
('i',each y)=: i.#y
)
Срд, 05 Сен 2012, Henry Rich писал(а):
> OK, I have fixed (or at least improved) detection of explicit
> definitions, and I now avoid processing nouns (as long as they are
> detected).
>
> I did not make the change you requested about removing the boxing.
> It's easy for a user to remove the boxing, and besides, it doesn't
> add much space, considering that if you have a lot of errors the
> character version will be padded up to the line-length of the longest
> error. But more important, any subsequent analysis tool would like
> to know the line number in error, so it makes sense to leave that in
> its own column.
>
> The fractional line numbers are used internally to sequence lines
> containing control words. I have removed them from the output now.
>
> Henry Rich
>
>
> On 9/5/2012 1:32 AM, bill lam wrote:
> >Can you change (or provide an option) the output to plain literal?
> >The box lines consume log of space and confuse parsing.
> >
> >also what is the meaning of decimal part of line numbers in
> >lint'jzplot'?
> >
> >|6503 |Sentence has no effect |
> >+---------+-----------------------------------------------------------------------+
> >|6503 |Syntax error - multiple assignment of non-noun after: start of
> >line |
> >+---------+-----------------------------------------------------------------------+
> >|6507.1667|Syntax error after: start of line invalid parts of speech |
> >+---------+-----------------------------------------------------------------------+
> >|6509 |Sentence has no effect |
> >+---------+-----------------------------------------------------------------------+
> >|6509 |Undefined name: Gxywh |
> >+---------+-----------------------------------------------------------------------+
> >|6515 |Undefined name: rounddist |
> >+---------+-----------------------------------------------------------------------+
> >|6521.3333|Undefined name: YCAPTION |
> >+---------+-----------------------------------------------------------------------+
> >|6521.3333|the test block must produce a noun value |
> >+---------+-----------------------------------------------------------------------+
> >|6522 |Undefined name: YCAPTION |
> >+---------+-----------------------------------------------------------------------+
> >|6523.3333|Syntax error after: start of line invalid parts of speech |
> >+---------+-----------------------------------------------------------------------+
> >|6526.3333|Syntax error after: start of line invalid parts of speech |
> >+---------+-----------------------------------------------------------------------+
> >|6530 |Sentence has no effect |
> >+---------+-----------------------------------------------------------------------+
> >
> >Срд, 05 Сен 2012, Henry Rich писал(а):
> >>OK, I have fixed all the problems reported.
> >>
> >>Ian,
> >>
> >> The script you pointed to has a line
> >>
> >>[: 0
> >>
> >> and a place where the name
> >>
> >>tte_r
> >>
> >>, which was created by 5!:0, has unknown type. Use lint directives
> >>to get a clean run.
> >>
> >>Bill,
> >>
> >> jzgrid and jzplot now run without crashing. They produce 1300 and
> >>1000 errors respectively, mostly names that are defined by a verb.
> >>The next step would be to use lint directives, and ordering of the
> >>code, to eliminate the errors. Seeing how many break. and continue.
> >>words are used (they caused lint trouble and they will cause a human
> >>analyzer trouble too!), this might be a worthwhile project - but not
> >>one I plan to undertake.
> >>
> >> Thanks to all for the bug reports, which uncovered about 6 program
> >>logic errors. If anyone reading has a program they want to try,
> >>bring it on - you might be surprised, as I have been, about what
> >>errors lurk in a supposedly-debugged program.
> >>
> >>Henry Rich
> >>
> >>On 9/4/2012 9:46 PM, bill lam wrote:
> >>>I also find lint failed for explicit noun that contain quote,
> >>>
> >>>---- a1.ijs
> >>>foo=: 0 : 0
> >>>I don't know
> >>>)
> >>>
> >>>will crash
> >>> lint 'a1.ijs'
> >>>|open quote
> >>>| I don't know
> >>>| ^
> >>>| estartx=.I._1~:(<0 0)&{::@(exppatt&rxmatch)@>
> >>>}:^:('NB.'-:3{.>@{:)&.;:&.>lines
> >>>
> >>>
> >>>Втр, 04 Сен 2012, Henry Rich писал(а):
> >>>>If you have something that makes lint crash, I'd like to see it (I'm
> >>>>working through plot now).
> >>>>
> >>>>You shouldn't expect every working script to fly through lint. In
> >>>>particular, global initialization inside a verb will require that you
> >>>>use lint directives to notify lint what's happening.
> >>>>
> >>>>But I have gotten some pretty long apps to go cleanly through after I
> >>>>told lint what was going on.
> >>>>
> >>>>Henry Rich
> >>>>
> >>>>On 9/4/2012 9:28 PM, Ian Clark wrote:
> >>>>>Yes, I've found a script or two which makes lint fall over...
> >>>>>
> >>>>>But IMO the facility is useful and important enough to deploy it now,
> >>>>>even in its "alpha" state (as I think Henry called it). Every addon
> >>>>>has bugs.
> >>>>>
> >>>>>On Wed, Sep 5, 2012 at 2:16 AM, bill lam <bbill....@gmail.com> wrote:
> >>>>>>I found lint incapable for checking more complex scripts. specifically
> >>>>>> lint 'jzplot' or lint' jzgrid' will raise strange error.
> >>>>>>
> >>>>>>Втр, 04 Сен 2012, Henry Rich писал(а):
> >>>>>>>OK, I changed the menu to make the Ctrl-1 key available (will not be
> >>>>>>>available for a little while, whenever Chris rebuilds the base).
> >>>>>>>
> >>>>>>>The lint menu becomes available for any window created AFTER you load
> >>>>>>>lint. So if you do
> >>>>>>>
> >>>>>>> load 'debug/lint'
> >>>>>>>
> >>>>>>>you'll have it for scripts opened thereafter. But I assumed that any
> >>>>>>>lint user would load it during startup, as you ended up doing.
> >>>>>>>
> >>>>>>>Yes, it's a shy facility! I was reluctant to make a change that
> >>>>>>>would surprise anybody. But I like your idea, which I will repeat as
> >>>>>>>
> >>>>>>>0. Make the menu line visible if ~addons/debug/lint/lint.ijs is found
> >>>>>>>1. Make the menu line enabled for IJS
> >>>>>>>2. If the menu line is selected, require 'debug/lint' before running
> >>>>>>>lint
> >>>>>>>
> >>>>>>>
> >>>>>>>I'm willing to do this, but not without general agreement. Does this
> >>>>>>>sound OK, guys?
> >>>>>>>
> >>>>>>>Henry Rich
> >>>>>>>
> >>>>>>>
> >>>>>>>
> >>>>>>>On 9/4/2012 8:32 PM, Ian Clark wrote:
> >>>>>>>>Very nice, Henry.
> >>>>>>>>
> >>>>>>>>May I offer some suggestions to gild the lily? Sorry if there's too
> >>>>>>>>much detail here. It's meant to clarify my meaning, rather than try to
> >>>>>>>>tell you what you'll know even better than I do.
> >>>>>>>>
> >>>>>>>>I'm running on a Macintosh, so what I say may not be quite the
> >>>>>>>>behavior under Windows, but here goes...
> >>>>>>>>
> >>>>>>>>This is a very "shy" facility, and IMO it will be easily overlooked by
> >>>>>>>>a casual user. I had to read the code in ~system/extras/util/jijs.ijs
> >>>>>>>>to find out how to make the said menu item appear.
> >>>>>>>>
> >>>>>>>>I was only able to do so by inserting the line:
> >>>>>>>>
> >>>>>>>>load 'debug/lint'
> >>>>>>>>
> >>>>>>>>in my startup.ijs. It didn't work to enter the above line in the
> >>>>>>>>session window.
> >>>>>>>>
> >>>>>>>>Even then, I could only see the new menu item: Run > Window and Syntax
> >>>>>>>>Check (Ctrl+1) when a script was opened in an IJS window and the
> >>>>>>>>window was uppermost (=selected).
> >>>>>>>>
> >>>>>>>>AFAICS the script jijs.ijs sets a flag "nolint" by detecting the
> >>>>>>>>presence of the verb: lint_lint_ .
> >>>>>>>>I suggest instead setting "nolint" by detecting the presence of the
> >>>>>>>>script: ~addons/debug/lint/lint.ijs .
> >>>>>>>>If the script is present, configure the Run menu to show "Window and
> >>>>>>>>Syntax Check" all the time, but grayed out unless an IJS window is
> >>>>>>>>open and selected.
> >>>>>>>>...In fact, more like the behavior of menu item "editlint": Edit >
> >>>>>>>>Format Script ⌘L
> >>>>>>>>
> >>>>>>>>If the menu item: Run > Window and Syntax Check is selected, only then
> >>>>>>>>load lint.ijs, so the _lint_ locale is only present if required. I
> >>>>>>>>guess this can be done by having the verb: "runlint" execute:
> >>>>>>>> require 'debug/lint'
> >>>>>>>>just before the "runimmx1" sentence.
> >>>>>>>>...Again, much as is done in verb: "lint" -the prettyprinter which
> >>>>>>>>runs when "editlint" is selected.
> >>>>>>>>
> >>>>>>>>And BTW the "runlint" entry in noun: JIJSMAC needs to be:
> >>>>>>>> menu runlint "Window and Syntax &Check" "Ctrl+1" "" "";
> >>>>>>>>instead of:
> >>>>>>>> menu runlint "Window and Syntax &Check (Ctrl+1)" "" "" "";
> >>>>>>>>This then shows correctly in the Run menu on the Mac as:
> >>>>>>>> Window and Syntax Check ⌘1
> >>>>>>>>and the shortcut ⌘1 (=Maccy name for hotkey) then works okay (at least
> >>>>>>>>it does for me).
> >>>>>>>>
> >>>>>>>>Ian
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>On Thu, Aug 30, 2012 at 11:57 PM, Henry Rich <henryhr...@nc.rr.com>
> >>>>>>>>wrote:
> >>>>>>>>>In the latest J602 base library, if you have lint installed, you
> >>>>>>>>>will get a
> >>>>>>>>>Run menu option to run lint on the current window. Ctrl-1 is the
> >>>>>>>>>shortcut
> >>>>>>>>>for this action.
> >>>>>>>>>
> >>>>>>>>>Henry Rich
> >>>>>>>>>----------------------------------------------------------------------
> >>>>>>>>>For information about J forums see
> >>>>>>>>>http://www.jsoftware.com/forums.htm
> >>>>>>>>----------------------------------------------------------------------
> >>>>>>>>For information about J forums see http://www.jsoftware.com/forums.htm
> >>>>>>>>
> >>>>>>>----------------------------------------------------------------------
> >>>>>>>For information about J forums see http://www.jsoftware.com/forums.htm
> >>>>>>
> >>>>>>--
> >>>>>>regards,
> >>>>>>====================================================
> >>>>>>GPG key 1024D/4434BAB3 2008-08-24
> >>>>>>gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3
> >>>>>>----------------------------------------------------------------------
> >>>>>>For information about J forums see http://www.jsoftware.com/forums.htm
> >>>>>----------------------------------------------------------------------
> >>>>>For information about J forums see http://www.jsoftware.com/forums.htm
> >>>>>
> >>>>----------------------------------------------------------------------
> >>>>For information about J forums see http://www.jsoftware.com/forums.htm
> >>>
> >>----------------------------------------------------------------------
> >>For information about J forums see http://www.jsoftware.com/forums.htm
> >
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
--
regards,
====================================================
GPG key 1024D/4434BAB3 2008-08-24
gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
